UNFAIR TRADE? EMPIRICAL EVIDENCE IN WORLD ...

26

UNFAIR TRADE?EMPIRICAL EVIDENCE IN

WORLD COMMODITY MARKETSOVER THE PAST 25 YEARS 1

Jacques Morisset

April 1997

1 I would like to thank Marcelo Olarreaga, Marc Bacchetta, Michael Finger, Neda Pirnia, Cheikh Kane,

Stijn Claessens, Joel Bergsman, Alejandro Izquierdo, and Antonio Estache for their valuable comments.These findings are my own and should not be attributed to the World Bank Group or its affiliates. Remaining errors are my responsibility. The address for correspondence is Foreign Investment AdvisoryService, 1818 H Street, NW, Washington, D.C. 20433, or (e-mail) [email protected]

27

TABLE OF CONTENTS

Page

Introduction....................................................................................................................3

I. Commodity Markets: Measuring the Variations in Spreads between World and Domestic Consumer Prices....................................................................4

II. The Asymmetric Response of Domestic Consumer Prices to Changes in World Prices....................................................................................8

III. How to Explain the Asymmetric Response of Domestic Prices...........................11

IV. What Are the Consequences for Commodity Exporting Countries?....................17

V. Concluding Remarks............................................................................................20

Bibliography.................................................................................................................24

Annexes........................................................................................................................26

28

Introduction

Since the 1970s, commodity prices have fallen in international markets. During

the same time, however, prices for consumers in industrial countries have risen. For

example, the price of coffee declined by 18 percent on world markets but increased by

240 percent for consumers in the United States between 1975 and 1993. Such diverging

patterns can be generalized across a wide sample of commodities and countries; from

crude oil to coffee; from Italy to the United States, but remain largely unexplored in the

current economic literature.

This paper looks at the spreads between international and domestic commodity

prices, then explains why these spreads have increased and analyzes their implications for

commodity exporting countries. The main finding is that the spreads have increased

dramatically because of the asymmetric response of domestic consumer prices to

movements in world prices. In all major consumer markets, decreases in world

commodity prices have been systematically much less transmitted than increases to

domestic consumer prices. This asymmetric response, which has been attributed to trade

restrictions and bidding processing costs, appears rather to be largely caused by the

behavior of international trading companies. The role of these companies merits greater

attention. While more evidence is still needed, I nevertheless show that many of these

companies are large enough to have a dominant position on most commodity markets.

Whatever the reason for the increasing spreads, their impact has been great: they may

have cost commodity exporting countries over US$100 billion a year because they have

limited the expansion of the final demand for these products in the major consumer

markets.

This paper argues that a special effort should therefore be made to understand the

determinants of the price of each of the consumer goods associated with commodities.

This effort should include the collection of information on international trading

companies, despite their general protectiveness, in order to improve transparency and

competition in these markets. Economists should also attempt to integrate intermediaries,

a subject that remains largely ignored by the mainstream literature, in the international

trade theory. Ultimately, only a better understanding of these companies will remove the

suspicion of unfair trade in international commodity markets.

29

The paper proceeds as follows. In the first section, empirical evidence on the

evolution of the spreads between world and domestic consumer prices is provided for

several commodities over the past 25 years. A discussion of the data used throughout the

paper is also included in this section. The second section is devoted to the relationship

between world and domestic prices using a time-series analysis. Special attention is

given to the asymmetric response of domestic prices to variations in world prices. The

explanations for this behavior range from trade restrictions to the role of international

trading companies, which are reviewed in the third section. The fourth section presents a

simple partial model that illustrates some of the potential negative implications arising

from the increase in the spreads over the past two decades. The last section contains

concluding remarks and possible directions for future research.

I. Commodity Markets: Measuring the Variations in Spreads between World and

Domestic Consumer Prices

Consumers in industrial markets can easily observe that prices of coffee, rice,

beef, and gasoline have increased almost continuously over the past two decades. When

these prices have declined, it has only been because of the short-term corrections to

episodes such as the oil price shocks in the 1970s. This generalized increase in consumer

prices can be contrasted with the declining long-term trend of world commodity prices;

for example, the World Bank’s non-fuel commodity index declined by 11 percent in

nominal dollars or 42 percent in constant dollars between 1980 and 1994.2 It is not

surprising, therefore, to find that the spread between the international and domestic

commodity prices increased dramatically during this period. This section shows, first,

how to measure the variations in these spreads and then gives the results for a sample of

commodities and countries over the period from 1970 to 1994.

The variations in the spread between world and domestic consumer prices can be

measured by the following standard equation (expressed in log-variations):

(1) ∆µij = ∆pij - ∆(ejp*i)

where ∆µij is the variation in the spread (or markup) associated with product i in country

j, pij the domestic consumer price of product i in country j, ej the nominal exchange rate

(dollar/local currency) in country j, and p*i the world price of commodity i. Domestic

consumer prices rather than producer prices are used to capture the final demand for these

2 Source: “Commodity Markets and the Developing Countries”, World Bank Quarterly, February 1996.

30

products. Equation (1) reflects the evolution of the spread over time, but it does not

provide information on its size at any given point in time. The variations in the spread

can be the result of multiple factors that will be reviewed in the following sections of this

paper.

This equation was applied to a sample of seven commodities: bananas, beef, crude

oil, coffee, rice, sugar, and wheat. These commodities were selected with several factors

in mind. One aim was to choose commodities that have as little processing as possible in

order to limit the influence of exogenous factors. Another goal was to provide variation

in terms of the types of products. For this reason, five of these commodities are produced

in both industrial and developing countries, while two are tropical products (coffee and

bananas). Only one mineral commodity (crude oil) was selected because it is hard to

match one specific final product with such mineral commodities. The eight following

pairs of commodities/consumer products were associated: bananas/bananas; beef/beef;

crude oil/fuel oil; crude oil/gasoline; coffee/coffee; sugar/sugar; wheat/bread; rice/rice.

The data on domestic consumer prices were compiled on an annual basis for the

six following countries: Canada, France, Germany, Italy, Japan, and the US. The choice

of an annual frequency primarily reflects the need to economize on data collection efforts.

All data were handcopied from government publications of these respective countries.

This sample was constrained by unequal access to comparable national sources for all

countries at a fairly desegregated level in the World Bank/International Monetary Fund

Library in Washington, D.C. (see Annex A). Nevertheless, these countries should capture

a large portion of worldwide consumption. In addition, the differences in their trade and

tax policies as well as their production structures should guarantee enough diversity for

the sample. International commodity prices were drawn from the World Bank data base

(see Annex A). Finally, the exchange rate for every country was defined as the annual

average rate reported in the IMF's International Financial Statistics.

The results show an unambiguous positive long-term trend in the spreads. For

presentation purposes, the results are reported in index values rather than in percentage

variations in Figure 1 and Tables 1a and 1b. The base year is 1990 for all variables

(1990=100). Figure 1 shows that the (arithmetic) average spread for all commodities

(and all countries) has followed a positive trend over the past two decades, with an

acceleration during the 1980s. To account for the annual volatility produced by seasonal

and climatic factors in commodity markets, the trend is best captured by the 5-year

moving average of the spread index., which doubled from a value of 51 to 117 between

1975 and 1994. The decline in the early 1970s is principally explained by the behavior of

oil prices since the average index, which excludes this commodity, actually increased

31

during this period. Finally, the recent reduction in the spread observed during the period

from 1993 to 1994 is principally explained by the sugar and coffee commodities, whose

prices fell dramatically.

The increasing trend in the spread is robust across countries and commodities.

The spreads surged in all industrial countries between 1975 and 1994, ranging from an

increase of 80 percent in the United States to almost 150 percent in Japan (Table 1a).

Among the European countries, the strongest increase was observed in Italy, followed by

France and Germany. Similarly, the spreads rose in all commodity markets, by

descending order from the coffee to the banana markets (Table 1b). Most spreads

declined in the first half of the 1970s due to unexpected commodity price booms, but they

more than recovered during the 1980s. As a result, only the spread for crude oil/gasoline

was still lower in 1994 than in the beginning of the 1970s. Finally, the secular increase in

the spreads is also demonstrated when

Figure 1 : Average Spread Index

0

20

40

60

80

100

120

140

1970

1972

1974

1976

1978

1980

1982

1984

1986

1988

1990

1992

1994

Years

1990

=100

All Commodities Excluding Oil All Commodities (5-year moving average)

32

33

the coverage period is extended to the 1960s, at least for countries where the data was

readily available (France, Italy, and the United States).

II. The Asymmetric Response of Domestic Consumer Prices to Changes in World

Prices

Why did the results presented above show a dramatic increase in the spread of

most commodity prices over the past two decades? The answer lies in the asymmetric

response of domestic consumer prices to changes in world prices. This section presents a

simple empirical model of the relationship between the variations in world and domestic

prices and then examines the asymmetry in this relationship for the sample of

commodities surveyed in this paper.

The model used in this section is based on the approach developed by Mundlack

and Larson (1992), and briefly summarized here. This model assumes that world prices

play a significant role in setting domestic consumer prices but that exporters can

discriminate prices by using their monopolistic power.3 As a result, the impact of world

prices on domestic prices is likely to vary across export destinations and commodities.

The model also predicts that domestic prices will be influenced by the nominal exchange

rate (ejt), labor costs (wjt), and the lagged domestic prices (pijt-1). Labor costs should

capture processing costs in the importing country4 (see explanation in the next section),

while the lagged dependent variable accounts for the presence of accumulated stocks and

fixed-in-advance contracts between buyers and sellers in most commodity markets (see

Anderson and Tyers [1992]). Other factors, such as changes in income in the destination

market, may also play a role, although most would be of secondary importance due to the

magnitude and variability of world commodity prices relative to changes in income.

Transportation costs, marketing costs, trade barriers, and health and safety regulations

that create subtle product differentiation were not introduced into the model due to the

lack of homogenous data. The influence of these factors will therefore be examined in

the next section.

The general model of domestic consumer price adjustment I propose to estimate

for the seven commodities in the six main consumer markets covered in this paper can be

written as follows:

3 This approach is similar to the one followed by the authors interested in the transmission of exchange

rate variations to domestic prices, the so-called “pass-through” literature. See Knetter (1993), for a goodsummary.

4 Labor costs were measured as the average unit labor cost in each industrial country covered in oursample. The data were extracted from the International Monetary Fund or UNIDO.

34

(2) ∆pijt = β∆p*it + γ ∆ejt + ρ∆wjt + φ∆pijt-1

All variables are defined in the text. The coefficient β is the elasticity of the

change in the domestic price with respect to the change in the world price, to be referred

to as the elasticity of transmission. The statistical interpretation of the β’s is

straightforward. A value of 1 implies that the variations in world prices are fully

transmitted to domestic prices. However, a perfect correlation should not be expected

since the commodity price is unlikely to account for 100 % of the consumer price. What I

try to show first is that there exists a significant and positive relationship between these

two prices and then, that this relationship is asymmetric. The above equation was

estimated for six countries and seven commodities from 1975 to 1994 using the random-

effect estimation technique (see detailed results in Annex B). Bananas and rice were

dropped because the data on their consumer prices were not available for all industrial

countries surveyed in this paper.

Overall, the estimated elasticities of transmission indicate a positive and

significant relationship between world and domestic prices in commodity markets (Table

2). The values of the elasticities are relatively low but such results can be expected with

regressions in variations rather than levels.5 A large portion of the price transmission

seems to be made within one year, in contradiction with the results found by Anderson

and Tyers for the 1960s and 1970s. The difference may be due to the more recent

coverage period used in this paper, for it reflects the emergence of the large commodity

funds in the 1980s, which have increased arbitrage opportunities and possibly shortened

the transmission time between world and domestic prices.6

So far, the model assumes that upward and downward movements in world

commodity prices have been equally transmitted to domestic prices. But, in reality, the

elasticity of transmission may differ in periods of increasing or decreasing world prices.

For example, the surge in oil price was almost perfectly passed on to domestic fuel prices

in the early 1970s, but the decline of 30 percent observed in the early 1990s was not

transmitted to domestic gasoline prices, which actually rose on average by 5 percent in

the six countries surveyed in this paper. More generally, the asymmetric response of

domestic prices was tested by estimating equation (2) for the years of increasing and for

those of decreasing world prices. The results for these two respective sub-periods are

5 I use variables in first differences to reduce the possibility of spurious correlations associated with time-series data when measured in levels.

35

presented in the “Upward Movements” and “Downward Movements” columns of Table

2.

Table 2:

Short-term and Long-term Elasticities of Transmission

Total Period Upward Downward

Short-Run Long-Run Movements a/ Movements a/

Coffee .25 .34 .31 .15

Sugar .03 .06 .15 -.04 *

Wheat .03* .05 .23 -.13

Beef .10 .11 .26 .12*

Gasoline .15 .15 .24 .17

Fuel .13 .14 .32 .16

Note: (*) not significantly different from 0 at a 5 percent

level.

a/ Only short-term elasticities are recorded because the

long-term elasticities cannot be estimated for upward and

downward movements due to the discontinuity of the years

analyzed.

The empirical results seem to support the hypothesis of asymmetric transmission

of movements in world prices in all commodity markets. The elasticity of transmission

has always been much higher, on average 3.4 times higher, when the world prices were

increasing rather than decreasing. Any decline in the international prices of sugar and

beef is unlikely to be passed on to consumer prices, while reductions in petroleum and

coffee prices are transmitted but much less than the corresponding increases. If upward

movements are perfectly transmitted but downward movements are not the spread

between world and domestic prices will increase continuously over time, as reported in

the first section of this paper. By comparison, Knetter [1993] found the inverse result for

a sample of manufacturing products. Prices adjusted more rapidly to exchange rate

6 For a study of the long-term relationship between world and domestic prices, a co-integrated approach

could be developed along the lines followed by Palaskas (1995). However, the limited number of annualobservations for each commodity prevented a similar approach in this paper.

36

depreciation (equivalent to a decline in world prices), suggesting that exporters of

manufactured goods choose to increase their market shares rather than their markups.

Similar behavior could not be shown in commodity markets.

Finally, the transmission from world to domestic prices has been remarkably

similar in all consuming countries surveyed in this paper. The elasticities of transmission

do not significantly differ across countries, as shown by the weak performance of the

fixed-effect technique.7 This finding was confirmed by the fact that the spreads of each

commodity moved jointly in all industrial countries. The cross-country contemporaneous

correlation between the spreads ranges from a minimum of 0.53 in the fuel market to a

maximum 0.95 in the gasoline market (Annex C).8 Since international effects appear to

be more important than host-country effects in explaining the asymmetric response of

domestic prices, the next section focuses exclusively on these effects.

III. How to Explain the Asymmetric Response of Domestic Prices

Explaining the growing spreads and the asymmetric price transmission is clearly a

matter of investigating the determinants of the price of each of the consumer goods in my

sample. One approach is to carefully examine each product in every country. The

quantity of data required is clearly beyond the scope of this paper. A second possibility

and the one I have selected follow a global approach that is, in my view, justified by the

homogeneity of the increasing spreads across countries and commodities.

There are multiple possible explanations for the asymmetric response of domestic

prices to changes in world commodity prices, which obviously, cannot occur in a

frictionless competitive model of trade. The two most popular explanations are the

presence of trade restrictions in the main consumer markets, and increasing processing

costs that act as bottlenecks in the trade of commodities. Still, these two explanations

seem to be a drastic simplification of the reality. While no consensus will emerge yet,

this section suggests that the market power of intermediaries, international trading

companies, is another possible explanation for the asymmetry. Surprisingly, their role

has been largely ignored in the economic literature. 9

7 Results are available upon request.8 Notice that, on the contrary, the variations in the spread of different commodities are only weakly

correlated within each country (see Annex C for a presentation of the contemporaneous correlation).9 The market power exerted by exporting countries is not considered in this paper. These countries can

influence world prices but certainly not their transmission to domestic consumer prices. The role of nationalmarketing boards and producers’ cartels is a different issue that clearly goes beyond the scope of this paper.

37

The first explanation is based on the existence of trade restrictions in most

industrial countries, and has been used by many authors interested in explaining the

asymmetric transmission of exchange rates (see Knetter [1993]). It suggests that in the

presence of binding quantity constraints in export markets, the decline in world

commodity prices will not be transmitted to domestic prices because there is no incentive

for exporters to stimulate the final demand by reducing their selling prices. Exporters

will instead increase their margins. Empirical support to this theory is provided by the

numerous import barriers faced by commodity exporters in consumer markets (see

Anderson and Tyers [1994] for examples). The asymmetric transmission of world

commodity prices has also been enhanced by using instruments specifically designed to

insulate domestic producers from lower world prices. Perhaps the most notorious

examples are the levies and variable tariffs adopted as part of the European agricultural

policy, but examples can be found in other industrial countries as well (see Mitchell and

Duncan [1987]).

The second explanation for the asymmetric response of domestic prices is that

exporters face a series of binding internal constraints when they want to increase their

sales abroad. For example, Foster and Baldwin [1986] introduce an approach using a

fixed proportion marketing technology that is required to sell products in the foreign

markets. This approach predicts that declines in world prices will be only imperfectly

transmitted to domestic prices because, if existing sales are constrained by marketing

capacity, exporters will compensate for increasing marketing costs by raising their selling

prices. This increase will partially offset the initial impact of declining world prices on

domestic prices. Since there is no similar constraint on higher world prices, one might

expect more domestic price adjustments to occur with rising than with declining world

prices. Potentially, this bottleneck approach can apply to a variety of costs, such as

processing, distribution, marketing, and transportation, all of which play a significant role

in setting domestic prices in commodity markets.

Table 3:Spreads and Effective Rates of Protection (ERPs)(Percentage change between 1986-88 and 1989-93)

Europe a/ Japan United States

Sugar ERP -38% -16% -49%Spread -13% -16% -34%

38

Wheat ERP -36% -24% 0%Spread 9% 1% 7%

Coffee ERP na na 0%Spread 23% 33% 45%

Beef ERP 17% -54% -33%Spread 7% 6% 6%

Rice ERP -33% -20% 100%Spread 6% -1% 4%

Sources: Ingco (1995) for the effective rates of

protection and my calculations for the spreads.

Notes:

a/ Only Germany, France, and Italy

The contribution of trade restrictions and bottleneck costs to the asymmetric

response of domestic prices might not be as important as appears at first sight. Indeed,

the variations in trade restrictions are weakly correlated to the movements in the spreads

for the commodities and countries surveyed in this paper. The weakness of this

correlation is most apparent when, despite significant differences in trade protection

between Europe, Japan, and North America, the spreads have moved almost

simultaneously in all these regions (see Annex C). The flaws of the hypothesized link are

further exposed by the weak correlation between the effective rates of protection and the

spreads.10 As reported in Table 3, only in the case of sugar did these two variables move

in the same direction in all consumer markets between 1986 and 1994. Finally, it is

certainly audacious to think that movements in trade barriers have significantly

contributed to the surge in the spreads of coffee and rice in the United States, up 85

percent and 112 percent, respectively, over the period from 1975 to 1994, when their

effective rates of protection were on average below 2 percent during this period.

Even the bottleneck approach does not work well for the simple reason that the

costs associated with commodity exports have been declining over the past few decades.

Indeed, transportation and insurance costs, which may contribute up to 10-20 percent of

10 Effective rates of protection present the advantage of capturing both the effects of both tariffs and non-

tariff barriers. Obtaining exact measurements of the effective rate of protection is always difficult, even forrelatively homogenous products such as foodstuffs. The differing qualities of products to which availableprice data refer and the presence of data on marketing margins are but two of the problems associated withusing even the simplest indicator of the extent of distortions.

39

the final value of commodities,11 have followed a descending trend over the past 20 years.

For example, Amadji and Yeats [1995] report that the share of these costs in the total

exports of developing countries declined from 7.8 percent in 1970 to 5.8 percent in 1991.

The international evidence on marketing and distribution costs is more limited, but the

trend in the United States has also been clearly negative,-down from 18 percent of GDP

in 1980 to only 10 percent of GDP in 1994.12 Technological progress and new

management techniques have clearly contributed to this trend. Among many examples,

electronic data interchanges have powered up market clearing activities, and just-in-time

techniques as well as new hedging instruments (e.g., warehouse bonds) have reduced

consignment and inventory costs.

The bottleneck approach may, however, partially explain the asymmetric

transmission of world commodity prices through rising processing costs, even though

their influence was limited by the kind of commodities selected in this paper. Unlike

transportation and marketing costs, processing costs have certainly increased over time

due to higher wages in processing facilities (most are located in industrial countries). The

direct evidence at hand remains sketchy but there is no reason to believe that these wages

have behaved differently from average industrial wages. And, over the past two decades,

average nominal industrial wages have seen a fivefold increase in the six countries

analyzed in this paper. Higher processing costs can also be explained by the improved

quality of consumer products such as unleaded gasoline and high-quality coffee (robusta

vs. arabica). Nevertheless, processing costs need to play a very important role in sales to

explain the asymmetric response of consumer prices. As an illustration, I estimated that

the impact of the average labor costs --as a proxy for processing costs-- on domestic

consumer prices should exceed by four times that of world prices to compensate entirely

for the increasing gap between world and consumer prices in the commodity markets

examined in this paper. 13

If the other explanations cannot provide a satisfactory answer to the rising

spreads, another reason has to be found. The third explanation for asymmetry is derived

from the presence of large trading companies in international commodity markets. The

focus is on the large trading companies because their strategic position between buyers

11 Atkin (1992) reports that transportation costs may account for 10 percent of the landed price of grainon a trade route between efficient ports used by large vessels (e.g., from New Orleans to Rotterdam) and 20percent on a less efficient route.

12 Source: Logistic Management Council (1996).13 In other terms, equation (1) was modified as follows: ∆µij = ∆pij - α∆(ejp*i) - (1-α)∆wj where wj is

defined as the unit labor cost in the recipient country j and α as the weight of the world commodity price inthe production function. The value of the parameter α is difficult to estimate in the absence of precise

40

and sellers allows them to influence the transmission of world prices. Such an effect may

occur when they purchase commodities from producers and/or when they sell these

products to other intermediaries, processors, and consumers. These companies generally

provide information, define the terms of transactions, manage the payments and record

keeping for transactions, and so figure out ways of clearing the market (see Spulber

[1996]). However, without competition, they may follow a pricing strategy that will

maximize their profits and not those of producers and consumers. Such behavior could

create an asymmetric response of the same sort as the bottleneck and trade restriction

models described earlier.14

The issue of the market power of international trading companies remains largely

ignored in the current literature. Several recent empirical studies have shown the

existence of market power in most commodity markets,15 but none of the leading

journals of international trade and economic development16 contain any reference to the

influence of these companies. This lack of interest possibly arises from the difficulty of

capturing the behavior of these companies in an integrated analytical framework. In

addition to their trading activities, many companies are vertically integrated and thus

close to production. For example, Cargill--the world’s largest trading company of

cereals--owns plantations, storage facilities, and vessels in many countries around the

world. Similarly, Exxon carries out not only mining and refining but also a complex set

of activities involving distribution, transportation, inventories, and pricing. The

distinction between wholesale and retail trading is also not clear-cut. If most of these

companies are involved in wholesales--transactions between business--there are many

examples in which they also act in the retail sector either directly or indirectly through

strategic alliances or intermediary arrangements.17 Additional studies are necessary to

identify at the stage of the intermediary process at which the highest profit is likely to be

made: wholesale or retail. The response is likely to vary across countries and

commodities.

information but must be as low as 0.2 for eliminating the spread between world and domestic prices in mostcommodity markets over the period from 1975 to 1994. These results are available upon request.

14 While it is not done in this paper, a model of imperfect competition --or price leadership-- behaviorcould show that declines in world prices will not be transmitted to consumer prices, and the output level willnot increase, at least not as much that in a competitive market. In contrast, an increase in world priceswould be automatically transmitted to domestic prices because intermediaries maintain their margins.

15 Recent studies include Buschena and Perloff (1991) on the coconut oil export market; Karp andPerloff (1989, 1993) on the rice and coffee exports; Lopez and Yon (1993) on the Haitian coffee exporting;and Deodhar and Skeldon (1995) on the banana export markets.

16 Sources examined (for the past five years) were the Journal of Development Economics and theJournal of International Economics as well as the NBER working paper series. Notice, however, that thisissue has been raised by non-mainstream economists such as Brown (1992).

17 For example, Itoh, the world’s largest wholesaler, owns coffee shops and pubs, and most oil companiespossess gas stations. Citgo, Texaco, Shell, Amocco, Exxon, and Chevron are the largest gasoline brands bynumber of stations, and are major wholesalers and distributors as well.

41

Table 4:

The World’s Largest Wholesale Trade

Companies: 1988

Firm Home

Country

Sales

(US$ Million)

C. Itoh. Ltd. Japan 106,791

Mitsui & Co. Ltd. Japan 102,493

Marubeni Corp. Japan 95,823

Sumitomo Corp. Japan 94,479

Mitsubishi Corp. Japan 91,583

Nissho Iwai Corp. Japan 52,942

Cargill US 43,000

Tokyo Menka Kaisha Japan 31,945

Sharps Pixley Ltd. UK 30,077

Nichimen Corp. Japan 26,874

Source: Directory of the World’s Largest Service

Companies, Moody’s Investors Service, and United

Nations Centre on Transnational Corporations,

December 1990.

Preliminary evidence indicates that large trading companies have been capable of

influencing the transmission of world commodity prices to domestic prices. This is

suggested first by the concentration of trading activities in few companies worldwide.

UNCTAD has reported that six or fewer trading companies control about 70 percent of

the total international trade, thus obviously limiting the choice of producers and

consumers in these markets.18 As an example, the banana export market is dominated by

Del Monte, United Brands, and Standard Fruits, and the wheat export market by Cargill,

Continental, Andre, Dreyfuss, and Bunge-Born. The suspicion that these companies use

their dominant position to control prices is strengthened by the chronic absence of

information on their activities. While many people can name retailers, few know

wholesalers. These companies are often larger than the economies of many developing

countries (Table 4). For instance, the sale volume of the world’s largest trading company,

C. Itoh, was as big as Argentina’s GDP in 1988. The same company also traded over

18 Source: UNCTAD, reported by Brown (1992).

42

US$20 billion of agricultural products--as much as all the sugar, coffee, beef, rice, and

wheat exported by all developing countries at that time.

The trading companies’ position of influence on the world market is further

implied by the correlation between the variations in the spreads and the variations in the

profits of the trading companies. Unfortunately, this hypothesis was tested only for the

oil market because of the chronic lack of data on these intermediary companies. For each

10 percent variation in the spread between world and domestic oil prices, the profit of the

7 largest oil companies in the United States has changed on average by 8 percent during

the period from 1979 to 1994.19 Another indicator of correlation is that the markup in the

wheat market grew by 50 percent over the past two decades, while the sales of Cargill, the

world’s largest trader of wheat, saw a fivefold increase during this period. In a historical

perspective, it is suggestive that this firm has recorded an annual loss in only 3 of its 130

years of existence: 1921, 1936, and 1938.20

Finally, as discussed in the preceding section, the spreads of each commodity tend

to move jointly in all industrial consumer markets. This homogenous behavior may

reflect the influence of trading companies that are specialized in trading one commodity

around the world rather than several commodities in one country. Companies such as

Cargill and Continental trade almost exclusively in cereals in over 60 countries. A

similar approach is taken by the petroleum trading companies and therefore gasoline

prices have a tendency to increase and decrease at the same time around the world.

IV. What Are the Consequences for Commodity Exporting Countries?

Rising spreads have had important consequences for commodity exporting

countries, especially for those depending heavily on a few commodities. Over the past

two decades, these countries have lost through the decline in world commodity prices and

through the limited response of domestic demand for these products on main consumer

markets. This section attempts to estimate how much additional export revenue these

countries would have earned if the spreads had remained constant in the past few years,

using a simple model of international trade. Finally, the results of two simulation

exercises are presented for the sample of commodities surveyed in this paper.

19 Calculated on the basis of information extracted from Fortune (various issues). To make the

measurement of profits and markups compatible, the profit is defined as the ratio of total net profits of largeUS oil companies to the international petroleum price (1990=100). The markup index is measured byequation (1). The major oil companies include Exxon, Mobil, Texaco, Chevron, Amoco, AtlanticRichfield, Philips Oil, and Ashland Oil.

20 Source: The Economist, March 1996.

43

The consequences of rising spreads on export revenues are illustrated as simply as

possible with a standard, partial model of international trade in which the commodity

supply function is determined by world prices and the demand by domestic prices in

consuming countries.21 For the sake of simplicity, these two functions are not influenced

by changes in relative prices and income, which are subsumed in the constant term of

these functions. There are neither dynamic effects nor strategic interactions between

trading companies as the variations in the spreads are assumed to be exogenously

determined. The model is principally intended to show the potential impact of rising

spreads rather than analyze actual pricing decisions. Nevertheless, it is easy to show that

lower spreads reduce domestic consumer prices, which increases the final demand for

commodities and, thus, export revenues. Obviously, the magnitude of these effects will

depend on the reduction in the spreads and the values of supply and demand price

elasticities.

The above model was applied to the sample of commodities over the period from

1991 to 1994. Rather than estimating the elasticity values of the demand and supply

functions, I used those estimated by the United Nations [1990], which are in the lower

range reported by Goldstein and Khan [1989]. These values are fixed over time, even

though they should vary as changes in prices imply changes in the degree of policy

intervention and in the degree of substitutability between products. However, within

feasible ranges, these variations should not modify the basic reliability of the results

presented below. The exogenous variations in the spreads are assumed to equal the

21 Thus, the demand and supply functions can be written as follows:

Qsi = A ep*iεs

Qdij = C pijεd

where εs and εd are defined as the elasticity of supply and demand, A and C as constantparameters, Qdij the demand for commodity i by consumers in country j, and Qsi the supply of commodity iby all developing countries. Other variables have been defined earlier.

Taking the log differential of the above equations and of the markup defined as µ = pi/p*, theeffects of a change in markup on export revenues (dRi) and producer surplus (dSi) are equal to:

dRi = - ((1+εs)εd)/(εd-εs)] dµi

dSi = (C/(εs+1)) [(1- εd/(εd+εs)dµi)ep*i)εs+1 - p*i

εs+1]

The positive effects of a decrease in markups are embodied in these two differential equations. Alower markup reduces the selling price on industrial markets. That, in turn, generates an increase in thefinal demand. The resulting effect would therefore be positive on both the export revenues and theproducer’s surplus. The magnitude of these potential positive effects depends partially on the percentagevariation in the markup and partially on the (absolute) value of the elasticities of demand and supply.

44

percentage difference, first of all, between the actual spread and the minimum spread

observed during the period from 1970 to 1994 (case A) and, second, between the actual

spread and the average spread observed during the period from 1970 to 1994 (case B).

All the parameters used for these simulations are summarized in Annex D.

Table 5 shows that developing countries would have doubled their export

revenues from 1991 to 1994 if the spreads had remained at their minimal levels of the

past two decades. . If the spreads had been maintained at their average levels, additional

export revenues would have reached US$40 billion per year, or about 27 percent of the

actual revenues from the six commodities selected in this paper. The potential gains for

producers would have also ranged from US$29 billion in case B to US$96 billion in case

A. These results only apply to developing countries. Indeed, industrial countries may

have benefited from asymmetry through higher tax revenues, higher value-added in their

processing facilities, and higher intermediary margins in their trading companies, even

though their consumers are clearly among the major losers. An estimate of the net

potential gains/losses for the industrial countries would need to take into account these

redistribution effects.

45

Table 5:Main Results of the Simulation Exercises

(US$ Billion)

Export ProducerRevenue Gains Surplus Gains

Case A Case B Case A Case B

Oil (fuel) 102.1 33.0 77.1 22.9Rice 1.9 1.0 1.5 0.7Sugar 8.7 1.9 7.4 1.4Coffee 9.1 3.9 8.3 3.2Beef 0.9 0.4 0.6 0.3Wheat 1.3 0.5 1.1 0.4TOTAL 124.0 40.6 96.0 29.0

Memo:Oil (gasoline) 59.7 19.7 39.8 13.1

Notes:

Case A: Percentage difference between the 1991-94

spread and the minimum spread observed during the

period 1970-94.

Case B: Percentage difference between the 1991-94

spread and the average markup observed during the

1970-94 period.

The simulation results indicate that petroleum would have accounted for about 80

percent of these additional potential gains since this commodity represents a large

proportion of the total exports from developing countries. Other commodities would

have also witnessed a significant increase in their export earnings. For example, the

revenues derived from coffee, sugar, beef, and wheat exports would have more than

doubled in case A, and increased in the range of 20-60 percent annually in case B. These

results are consistent with the large percentage differences in the spreads observed for

these commodities.

As expected the developing countries that have suffered the most are those that

are heavily dependent on oil exports such as Saudi Arabia, the CIS countries, and Nigeria

(Table 6). Brazil is also a major loser due to its significant dependence on coffee and

46

sugar exports. For smaller countries, the consequences are even more dramatic because

they rely on only one or two commodities for their exports. For example, Mauritius may

have increased its total export revenues by an estimated 30 percent if the spread in the

sugar market had remained at its minimal level. Similar results are obtained in the coffee

market for El Salvador, Kenya, Madagascar, and Colombia (respectively, 50, 28, 27, and

25 percent of their total export revenues). The above results are only indicative. As

already mentioned, the model is extremely simple.

V. Concluding Remarks

The relatively low income and price elasticities of demand for commodities was

emphasized by Prebisch and Singer about 35 years ago. This paper goes one step further

by suggesting that the final demand for these products could not have increased in the

major consumer markets because the declines in world commodity prices were not

transmitted or were transmitted imperfectly to domestic consumer prices. In contrast,

upward movements in world prices were clearly passed on to domestic prices. As a result

of this asymmetry, the spread between world commodity prices and domestic consumer

prices has increased over time, about 100 percent on average for the seven commodities

analyzed in this paper over the past 25 years. This asymmetry has had severe

implications for the commodity exporting countries, who may have lost as much as

US$100 billion per year in export revenues.

In this paper, I have to attempted to review a number of possible explanations for

the asymmetry, which is the most logical way to proceed without an existing general

analytical framework in the economic literature. A consistent finding across commodity

markets has been the simultaneous movement of the spreads in all countries, thus

suggesting the influence of international rather than country-specific factors. There are at

least two international factors that may explain the asymmetric response of domestic

prices in commodity markets. First, the high quantitative restrictions on international

commodity trade have discouraged exporters from stimulating the final demand by

transmitting the decrease in world prices to domestic consumer prices. Second, the

processing costs have been increasing due to rising labor costs and improvements in the

quality of the final products associated with most commodities. In contrast, other costs

such as transportation, insurance, distribution, and

47

48

marketing do not appear to play a role in the rising spreads. These costs have followed a

declining trend over the past few decades and would thus explain a decline rather than an

increase in the spreads.

There is little consensus on this issue, but the above explanations do not seem to

provide a complete answer. Indeed, it appears that trade restrictions are only weakly

correlated with the movements in the spreads, an observation that is consistent across

countries and in one country over time. The contribution of processing costs to the

increasing spreads is certainly limited in most cases examined here because the sample of

commodities covered in this paper involves little processing between the commodity and

the final product sold on consumer markets. For these reasons, another explanation had to

be found to explain the asymmetric transmission of world prices.

This paper has argued that international trading companies are likely to influence

the relationship between world and domestic prices. Their dominant position in most

commodity markets enables them to affect the spreads between the buyer and the seller

prices simultaneously in many countries. Some preliminary evidence points in that

direction, but surprisingly policy-makers, economists, and consumers seem to remain

largely unaware of these companies, even though they are often bigger than developing

economies. The current academic literature as well as international institutions have

traditionally ignored their presence. This insufficient attention partially explains why the

debate over these companies lacks focus and clarity and why there are various

misconceptions about what these companies actually do and whether their activities are a

legitimate cause for public concern.

This paper should be viewed as a starting point for discussion. Possible directions

for future research include an attempt to better understand the determinants of the

consumer prices and of the role of intermediaries at both the wholesale and retail levels.

In that sense, the first recommendation would be therefore to collect information on the

activities of these companies. Competitive (or contestable) markets assume homogenous

information. Today, producers and consumers generally have few alternatives when they

trade their products in foreign markets because of the lack of information. Collecting

information will require a concerted effort from the international community. First, it is

crucial that the large international trading companies cooperate and disclose information

on their activities and transactions. Second, this effort must necessarily involve the

World Bank and the World Trade Organization because they have both the necessary

financial and human resources to undertake such an operation on a worldwide basis.

49

The second recommendation is that economists incorporate the subject of

intermediation within the basic framework of international trade. So far, trading

companies might have been overlooked because they are located at the crossroads of

different aspects of economic theory: business, industrial organization, international trade

and finance, as well as public finance. The new international trade theory has emphasized

the increasing rate of returns and imperfect competition but not at the intermediary level.

There is a need to understand the behavior of the trading companies as well as the

determinants of their pricing strategies to evaluate whether they operate efficiently. The

remaining issue is to determine whether these companies seek to maximize their profits at

the expense of those of consumers and producers.

Free trade requires fair trade. For the first time, anything can be sold everywhere

and thus understanding the role of the international trading companies in commodities

markets will become even more important in the future.

50

Bibliography

P. C. Abott (1988), “Estimating US Agricultural Exports Demand Elasticities:

Econometric and Economic Issues”, in International Agricultural Trade, C. Carter

(ed.), Boulder, Westview Press.

A. Amjadi and A. Yeats (1995), “Have Transport Costs Contributed to the Relative

Decline of Sub-Saharan African Exports?”, Policy Research Paper N.1559, The

World Bank, December.

M. Atkin, “The International Grain Trade”, Woodhead Publishing Limited, Cambridge,

1992.

K. Anderson and R. Tyers (1992), “Disarray in World Food Markets”, New York,

Cambridge University Press.

B. Balassa (1989), “Outward Orientation”, in Handbook of Development Economics, H.

Chenery and T. N. Srinivasan (eds.), North Holland.

M. B. Brown (1992),”Fair Trade”, London, Zed Book.

S. Deodhar and I. Sheldon (1995), “Is Foreign Trade (Im)perfectly Competitive? An

Analysis of the German Market for Banana Imports”, Journal of Agricultural

Economics, 46(3).

J. De Melo, M. Olarreaga, and W. Takacs (1996), “Pricing Policy Under Double Market

Power: Madagascar and the International Vanilla Market”, Working Paper,

Department of Economics, University of Geneva.

The Economist, “How to Feed a Growing Family”, March 9, 1996

J.M. Finger and U. Reinke (1995), ”The Uruguay Round and the Developing Economies:

Tariff Concessions Given and Received”, International Trade Division, The

World Bank.

M. Goldstein and M.S. Khan, “Income and Price Effects in Foreign Trade “, Handbook of

International Economics,

M. Ingco (1995), “Agricultural Trade Liberalization in the Uruguay Round: One Step

Forward, One Step Back”, Policy Research Paper N.1500, The World Bank,

August.

L. Karp & J.M. Perloff (1989), “ Oligopoly in the Rice Export Market”, Review of

Economics and Statistics, 71.

L. Karp & J.M. Perloff (1993), “A Dynamic Model of Oligopoly in the Coffee Export

Market”, American Journal of Agricultural Economics.

M. Knetter (1993), “International Comparisons of Pricing-to-Market Behavior”,

American Economic Review, vol. 83, June.

M. Knetter (1992), “Is Price Adjustment Asymmetric? Evaluating the Market Share and

Marketing Bottlenecks Hypothesis”, NBER Working paper, N. 4170, September..

51

R.E. Lopez & Z. You (1993), “Determinants of Oligopolistic Power: The Haitian Coffee

Case”, Journal of Development Economics, 41.

Moody Investors Service and United Nations (1990), “Directory of the World’s Largest

Service Companies”, Serie I, December.

D. Mitchell & R. Duncan (1987), “Market Behavior of Grain Exporters”, World Bank

Research Observer, January.

D. Morgan, (1979) “Merchants of Grain”, New York, Viking Press.

Y. Mundlack and D. Larson (1992), “On the Transmission of World Agricultural Prices”,

The World Bank Economic Review, September.

T. Palaskas (1995), “Statistical Analysis of Price Transmission in the European Union”,Journal of Agricultural Economics, 46 (1).

D. Spulber (1996), “Microstructure and Intermediation”, Journal of EconomicPerspectives, vol.10, n.3, Summer.

P. Tyers and K. Anderson (1989), “Price Elasticities in International Food Trade”, Journal

of Policy Modeling.

United Nations (1990), “Agricultural Trade Liberalization in the Uruguay Round”.

The World Bank (1995a), “The Uruguay Round: Widening and Deepening the WorldTrading System”, Direction in Development.

The World Bank (1995b), “Managing Commodity: Booms and Busts”, Direction in

Development.

52

ANNEX A:Data Sources and definitions

A. Description of Domestic Price Series a/

Commodity/End-UserProduct

Canada

France Germany

Italy Japan

USA

Bananas/Bananas

✔ ✔

Beef/Beef ✔ ✔ ✔ ✔ ✔ ✔

Oil/ Fuel ✔ ✔ b/ ✔ ✔ ✔ ✔

Oil/Gasoline ✔ ✔ ✔ ✔ ✔ ✔

Coffee/Coffee ✔ ✔ ✔ ✔ ✔ ✔ c/Rice/Rice ✔ ✔ ✔ ✔ ✔ d/Wheat/Bread ✔ ✔ ✔ ✔ ✔ ✔

Sugar/Sugar ✔ ✔ ✔ ✔ ✔ ✔ e/

Sources: National statistics for consumer price indexes and World Bank forcommodity price index.Notes:a/ The annual domestic consumer price series were available for thefollowing periods: Canada (1970 and 1975-94), France (1964-94), Germany(1966-94), Italy (1960-94), Japan (1973-94), and the US (1960-94).b/ Only available for the period 1971-94.c/ Only available for the period 1969-94.d/ Only available for the period 1978-94.e/ Only available for the period 1970-94.

B. Description of International Commodity Prices

Coffee: All Coffee, New York, US cents/LBSugar: Caribbean, New York, US cents/LB

53

Beef:, All origins, US Ports, US cents/LBWheat: US, US Gulf Ports, US$/BushelCrude Oil (petroleum): Average Crude Price, US$/Barrel:Bananas: Latin America, US Ports; US cents/LBRice: US, New Orleans, US$/MT

Source: The World Bank. International Economic Department

54

ANNEX B:Regression Results

Elasticity of Transmission from World Prices to Domestic Consumer Prices Panel of six countries (1975-94)

Coffee Beef Sugar (1) (2) (3) (1) (2) (3) (1) (2) (3)

World Price 0.25 0.31 0.15 0.10 0.26 0.12 0.03 0.15 -0.04(6.85) (5.51) (2.06) (2.30) (4.13) (.71) (2.26) (4.81) (-0.54)

Exchange Rate -0.02 0.02 -0.02 0.14 0.24 0.09 0.19 0.15 0.22(-.12) (.10) (-0.11) (2.38) (2.47) (1.23) (3.19) (2.04) (2.40)

Industrial Wage 0.44 1.13 0.01 0.21 0.17 0.77 0.49 0.24 0.40(2.05) (2.88) (.439) (2.01) (1.33) (5.54) (7.62) (2.56) (2.75)

Lagged Domestic Price 0.26 0.09 0.50(2.01) (1.97) (2.34)

AdjR2 0.32 0.43 0.05 0.13 0.25 0.24 0.15 0.33 0.17DW 1.93 2.24 2.33 1.63 1.86 1.92 1.48 1.72 1.73Observations 114 60 54 114 54 60 114 60 54

Wheat Oil/Gasoline Oil/Fuel (1) (2) (3) (1) (2) (3) (1) (2) (3)

World Price 0.03 0.23 -0.13 0.15 0.24 0.17 0.13 0.32 0.16(1.04) (3.08) (-2.10) (4.50) (4.32) (3.13) (2.98) (4.33 (1.64)

Exchange Rate 0.15 0.24 0.17 0.27 0.09 0.29 0.33 0.05 0.30(3.30) (2.29) (3.41) (3.18) (.82) (1.56) (2.59) (3.18) (1.00)

Industrial Wage 0.32 0.41 0.58 0.49 1.04 0.30 0.62 1.42 0.46(4.05) (3.82) (5.84) (5.00) (5.45) (1.87) (4.23) (5.48) (1.78)

Lagged Domestic Price 0.40 0.01 0.07(1.77) (2.13) (2.41)

AdjR2 0.23 0.12 0.39 0.29 0.50 0.01 0.22 0.48 0.01DW 1.62 1.41 1.71 2.18 1.80 2.13 2.17 1.81 2.28Observations 114 48 66 114 66 48 114 66 48

Notes:All variables are expresed in log and in variations.Column (1) are the estimated results for the entire period.Column (2) are the estimated results for the years with upward movements in world prices.Column (3) are the estimated results for the years with downward movements in world prices.

1

AnnexC:Contemporaneous Correlations

2

Cross-Country Correlation by Commodity1970-94

COFFEEJapan France Germany Canada Italy US

Japan 1.0

France 0.7 1.0Germany 0.5 0.8 1.0

Canada 0.9 0.5 0.6 1.0Italy 0.9 0.6 0.5 0.8 1.0US 0.9 0.2 0.0 0.8 0.8 1.0

FUELJapan France Germany Canada Italy US

Japan 1.0

France 0.7 1.0Germany 0.7 0.7 1.0

Canada 0.7 0.7 0.5 1.0Italy 0.5 0.0 0.3 0.3 1.0US 0.7 0.8 0.6 0.7 0.0 1.0

GASOLINEJapan France Germany Canada Italy US

Japan 1.0

France 1.0 1.0Germany 0.9 0.9 1.0Canada 0.9 1.0 0.9 1.0

Italy 1.0 1.0 1.0 1.0 1.0US 0.9 1.0 0.9 0.9 1.0 1.0

RICEJapan France Germany Canada Italy US

Japan 1.0

France 0.9 1.0Germany NA NA NACanada 0.8 0.8 NA 1.0

Italy 0.9 0.7 NA 0.8 1.0US 0.8 0.7 NA 0.9 0.9 1.0

WHEATJapan France Germany Canada Italy US

Japan 1.0

France 1.0 1.0Germany 1.0 1.0 1.0Canada 0.9 0.9 0.9 1.0

Italy 1.0 1.0 0.9 0.9 1.0

US 1.0 0.9 0.9 0.9 0.9 1.0SUGAR

Japan France Germany Canada Italy USJapan 1.0France 0.9 1.0Germany 0.9 0.9 1.0Canada 0.9 0.9 0.9 1.0Italy 1.0 1.0 0.9 1.0 1.0US 0.9 0.9 0.9 1.0 0.9 1.0BEEF

Japan France Germany Canada Italy USJapan 1.0France 0.9 1.0Germany 0.5 0.6 1.0Canada 0.8 0.8 0.3 1.0Italy 0.8 0.8 0.3 0.7 1.0US 0.4 0.6 0.4 0.5 0.5 1.0

3

Cross-Commodity Correlation by Country1970-94

USCoffee Banana Sugar Rice Bread Gasoline Fuel Beef

Coffee 1.0Banana 0.2 1.0Sugar 0.5 0.1 1.0Rice 0.8 0.3 0.6 1.0Bread 0.5 0.3 0.4 0.6 1.0Gasoline -0.7 -0.1 -0.5 -0.7 0.1 1.0Fuel -0.3 0.0 -0.3 -0.3 0.5 0.9 1.0Beef -0.3 -0.2 -0.2 -0.1 0.1 0.3 0.3 1.0JAPAN

Beef Banana Coffee Fuel Gasoline Sugar Bread RiceBeef 1.0Banana 0.5 1.0coffee 0.7 0.5 1.0Fuel 0.7 0.6 0.5 1.0Gasoline 0.5 0.6 0.4 0.7 1.0Sugar 0.5 0.5 0.5 0.5 0.2 1.0Bread 0.9 0.5 0.8 0.7 0.4 0.7 1.0

Rice 0.8 0.5 0.8 0.7 0.4 0.7 1.0 1.0ITALY

Bread Beef Sugar Coffee Fuel Gasoline RiceBread 1.0Beef 0.8 1.0Sugar 0.6 0.4 1.0Coffee 0.8 0.6 0.5 1.0Fuel 0.0 0.0 -0.4 0.0 1.0Gasoline 0.0 -0.2 0.2 0.3 0.2 1.0Rice 0.9 0.7 0.7 0.6 -0.2 -0.2 1.0GERMANY

Bread Sugar Fuel Gasoline Coffee BeefBread 1.0Sugar 0.1 1.0Fuel 0.0 0.2 1.0Gasoline -0.2 0.3 0.8 1.0Coffee -0.2 0.3 0.4 0.6 1.0Beef 0.3 -0.1 0.1 -0.1 -0.1 1.0FRANCE

Bread Beef Rice Sugar Coffee Gasoline FuelBread 1.0Beef 0.7 1.0Rice 0.7 0.8 1.0Sugar 0.5 0.4 0.6 1.0Coffee -0.1 0.3 0.4 0.4 1.0Gasoline -0.2 0.2 0.3 0.4 0.8 1.0Fuel 0.1 0.3 0.5 0.6 0.7 0.9 1.0CANADA

Beef Bread Rice Sugar Coffee Fuel GasolineBeef 1.0Bread 0.8 1.0

Rice 0.8 0.9 1.0Sugar 0.6 0.7 0.8 1.0Coffee 0.7 0.8 0.8 0.6 1.0Fuel 0.8 0.8 0.8 0.8 0.7 1.0Gasoline 0.3 0.4 0.5 0.3 0.5 0.6 1.0

4